An explosion precursor intelligent pre-judgment trigger control method of an active bidirectional explosion-proof valve

Abstract

The application discloses an explosion precursor intelligent pre-judgment trigger control method of an active bidirectional explosion-proof valve, and belongs to the technical field of industrial explosion-proof safety. The application aims at the core defects of the existing explosion-proof valve, such as response lag, one-way protection limitation, difficulty in identifying early explosion precursor signals, high misjudgment and missed judgment rate, and lack of collaborative control of bidirectional valves, and solves the problems by the following steps: through sub-microsecond multi-sensor time synchronization and phase calibration, and self-adaptive chaos system weak precursor signal enhancement, a multi-dimensional feature coupling solution model and a dynamic self-adaptive three-level pre-judgment threshold are constructed, a bidirectional explosion-proof valve collaborative trigger timing matching mechanism is designed, and a model closed-loop self-correction algorithm is combined. The application can realize the early pre-judgment and active trigger of 10-50 ms before explosion, greatly reduces the feature solution error, improves the bidirectional explosion-proof protection reliability, and is suitable for the bidirectional active protection of combustible gas and dust explosion in chemical, grain, tobacco, wood processing and mine scenes.

Description

Technical Field

[0001] This invention relates to the field of industrial explosion-proof safety technology, and in particular to an intelligent prediction and triggering control method for an active bidirectional explosion-proof valve. Background Technology

[0002] Currently, in industrial settings such as chemical processing, grain processing, mining, and powder conveying, explosions of flammable gases and dust within pipelines pose significant safety risks. Explosion-proof valves are core equipment for blocking explosion propagation and preventing the escalation of accidents. However, existing explosion-proof valve and control technologies suffer from the following inherent limitations: (1) Passive response lag, inherent defects in protection: The existing mainstream explosion-proof valves are passive structures that rely on the explosion shock wave to push the valve disc to close. They can only act after the explosion occurs, and the response time is usually >50ms. They cannot block the explosion precursor and the initial shock wave, which can easily cause damage to downstream equipment and cannot achieve inherent safety protection.

[0003] (2) The triggering logic is simple, and the false judgment and missed judgment rate is high: Most of the existing active explosion-proof valves use a fixed threshold trigger of a single pressure / flame sensor, which can only identify strong characteristic signals after the explosion occurs and cannot capture the precursors of the explosion; at the same time, the fixed threshold cannot adapt to the normal operating condition fluctuations in the pipeline (such as pressure pulsation, turbulence, and normal changes in medium concentration), and it is easy to be falsely triggered when the operating condition fluctuates and to be missed when the operating condition changes. The false judgment rate in industrial field measurements is generally >10%.

[0004] (3) Insufficient sensing synchronization accuracy and serious distortion of feature calculation: Existing sensing systems mostly use hard triggering or ordinary network synchronization. The time synchronization error of multiple sensors is in the millisecond level. At the same time, the inherent response delay differences of different types of sensors are not considered, resulting in calculation errors of core parameters such as explosion wave propagation speed and spatial gradient >30%, which directly causes prediction timing deviation and triggering timing error.

[0005] (4) Early weak signals cannot be identified, and the prediction ability is seriously insufficient: the early precursor signals before the explosion are mostly weak pulsations at the microvolt / nanovolt level, which are completely submerged in strong background noise such as fan vibration and medium turbulence. Existing conventional filtering technology will filter out the weak precursor signals and noise simultaneously, and cannot achieve effective extraction. As a result, the existing technology can only identify them after the explosion occurs and cannot achieve advanced prediction. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent prediction and triggering control method for explosion precursors in an active bidirectional explosion-proof valve, which can solve industry pain points such as response lag, large time synchronization errors of multiple sensors, and the inability to extract weak early-stage explosion precursor signals due to strong operating noise.

[0007] To achieve the above objectives, the present invention provides an intelligent prediction and triggering control method for pre-explosion signs of an active bidirectional explosion-proof valve, comprising the following steps: S1. Multi-dimensional sensing units are deployed on the forward and reverse pipeline sides of the explosion-proof valve. A sub-microsecond time synchronization network of all sensing nodes is constructed with the edge controller as the main clock node to complete the calibration of the inherent response delay of the sensor, the calibration of the baseline working condition characteristics, and the combined weight calculation of the characteristic parameters. S2. Real-time acquisition of sensing data, completion of phase calibration and adaptive chaotic amplification feature enhancement of weak early warning signals of explosion, calculation of single-parameter multi-dimensional anomaly degree, and then calculation of the explosion precursor coupling hazard degree of forward and reverse pipelines. S3. Calculate the real-time operating condition fluctuation coefficient, generate a three-level dynamic adaptive prediction threshold, and complete the graded prediction of explosion risk. S4. Calculate the real-time propagation speed of the explosion wave and the optimal triggering time of the bidirectional valve, and execute differentiated collaborative triggering control and adaptive control of valve disc closing speed. S5. Emergency closed-loop control is executed based on the feedback of the explosion-proof effect, and the adaptive self-correction of the prediction model is completed through iterative algorithm; S6. Upload the predicted and triggered data to the industrial internet platform in real time, link the factory emergency system, and complete the data storage and traceability of the entire process.

[0008] Preferably, in S1, all sensing units and edge controllers are networked via the IEEE 1588 PTP V2 Precision Time Protocol to achieve sub-microsecond time synchronization within ±100ns for all sensing nodes; the inherent response delay of each type of sensor is calibrated using a standard step signal generator; the combined weights of the feature parameters are calculated using the following formula: ; In the formula, For the first i The final combined weights of the feature parameters satisfy the following conditions: ; For the first i The subjective weights of each feature parameter are determined by the analytic hierarchy process (AHP). For the first i The objective weights of each feature parameter are calculated using the entropy weight method; This represents the total number of precursor characteristic parameters.

[0009] Preferably, in step S2, phase calibration is first performed on the acquired raw signal, using the following calibration formula: ; In the formula, for t Time of the first m The first sensing unit iThe effective signal after phase calibration of each parameter; These are the original values ​​collected by the sensor; For the first i The inherent response delay calibration value of this type of sensor; The real-time time deviation between the slave clock and the master clock of the corresponding sensing unit is positive when the slave clock lags behind the master clock and negative when it leads.

[0010] Preferably, in S2, the calibrated signal is enhanced using a Dove-Holmes chaotic system with adaptive periodic driving force to enhance the weak precursor signal characteristics. The system dynamic equation is as follows: ; In the formula, The system damping coefficient is... , The restoring force coefficient is nonlinear. As a periodic reference driving force, For the driving force amplitude, The driving force angular frequency, For signal input gain, Background Gaussian white noise term; By adaptively adjusting the amplitude of the driving force F By operating the system in a chaotic critical state, selective amplification and noise suppression of precursory weak signals are achieved, resulting in the output of enhanced characteristic signals. .

[0011] Preferably, in S2, the single-parameter multidimensional anomaly degree is calculated using the following formula: ; In the formula, for t Time of the first m The first sensing unit i Real-time anomaly degree of each feature parameter; The rate of change over time, The axial spatial gradient; , The first i The baseline mean and standard deviation of each parameter; , These are the benchmark mean and standard deviation of the rate of change over time, respectively. , These are the baseline mean and standard deviation of the spatial gradient, respectively. , , The weights are the absolute value of the parameter, the temporal gradient, and the spatial gradient, respectively, satisfying... + + =1.

[0012] Preferably, in S2, the formula for calculating the pre-explosion coupling hazard of the forward pipeline is: ; The formula for calculating the pre-explosion coupling hazard of a reverse pipeline is: ; In the formula, , The explosion precursor coupling hazard levels are respectively for forward and reverse pipelines; , These represent the number of forward and reverse sensing units, respectively. and The calculation method and same, For the first positive pipeline n The first sensing unit, the first i Real-time anomaly degree of each parameter For the reverse pipeline n The first sensing unit, the first i Real-time anomaly degree of each parameter; , The first m , n The positional weight of each sensing unit is greater the closer it is to the explosion-proof valve; , These are the propagation characteristic coefficients of the forward and reverse pipeline signals, respectively. The coefficient is 1 when the abnormal signal propagates to the explosion-proof valve, and 0.5 otherwise.

[0013] Preferably, in S3, the real-time operating condition fluctuation coefficient is calculated using the following formula: ; In the formula, For real-time operating condition fluctuation coefficient, satisfying ≥1; For the first time window within the sliding time window i Real-time standard deviation of each parameter; The three-level dynamic adaptive prediction threshold is calculated using the following formula: ; In the formula, , , These are dynamic thresholds for early warning level, pre-trigger level, and emergency trigger level, respectively. , , As the baseline threshold, satisfying 0 < < < ; when ≥ or ≥ When the warning level is triggered, ≥ or ≥ The pre-trigger level is triggered when ≥ or ≥ Emergency trigger level is triggered at this time.

[0014] Preferably, in S4, the real-time propagation speed of the explosion wave is calculated using the following formula: ; In the formula, , These represent the propagation speeds of the forward and reverse explosion waves, respectively. The axial spacing between adjacent sensing units; , These represent the propagation time differences between adjacent sensing units for positive and negative abnormal signals, respectively. The optimal trigger times for the forward and reverse explosion-proof valves are calculated using the following formulas: ; ; In the formula, , These are the optimal trigger times for both forward and reverse directions; The current moment; , These are the distances from the leading edge of the abnormal signal to the valve port; The rated full closing time for the explosion-proof valve.

[0015] Preferably, in S4, the valve closing speed is adaptively controlled by the following formula: ; In the formula, This refers to the real-time closing speed of the valve disc. This refers to the maximum rated closing speed of the valve disc. The coupling hazard level of the corresponding side pipe.

[0016] Preferably, in S5, the self-correcting iterative formula for the prediction model's weights is: ; In the formula, The combined weights are updated iteratively; The model learning rate; To predict errors, a value of 1 is used for missed predictions, -1 for incorrect predictions, and 0 for accurate predictions. For the firsti The real-time average anomaly of each parameter; the weights after iteration must satisfy upper and lower bound constraints and undergo normalization to ensure... .

[0017] Therefore, the beneficial effects of the above-mentioned intelligent prediction and triggering control method for the explosion precursor of an active bidirectional explosion-proof valve are as follows: (1) This invention breaks through the inherent mode of "triggered after explosion" in the prior art. By identifying the pre-explosion multi-physics field coupling, it can complete the advance prediction and complete valve closure 10-50ms before the explosion occurs, blocking the propagation of the explosion wave and the precursor from the root, and solving the essential defect of the passive valve response lag.

[0018] (2) This invention achieves sub-microsecond time synchronization within ±100ns for all sensing nodes through the IEEE 1588 PTP protocol. Combined with phase calibration of the inherent response delay of the sensor, it eliminates the calculation distortion caused by multi-sensor time synchronization error and phase deviation, reduces the calculation error of core parameters such as explosion wave propagation speed and spatial gradient, and provides core support for accurate prediction and triggering.

[0019] (3) The present invention uses the Dufen chaotic system with adaptive periodic driving force, which can extract the early precursor signal of the nanovolt level explosion submerged in strong operating noise, solves the problem of missed judgment caused by the inability of existing technology to identify early weak signals, and is conducive to improving the accuracy of prediction response.

[0020] (4) The present invention designs a dynamic threshold that is adjusted in real time according to the working conditions. Combined with the weight self-correction iterative algorithm, it can adaptively adapt to the changes in pipeline medium, working conditions and equipment characteristics throughout the entire life cycle. It does not require repeated manual calibration, and the long-term operating accuracy does not decrease. The equipment maintenance cycle is extended by more than 3 times.

[0021] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the steps in an embodiment of the intelligent prediction and triggering control method for an active bidirectional explosion-proof valve according to the present invention. Figure 2 This is a schematic diagram of the adaptive chaotic amplification and feature enhancement process for weak signals before an explosion, as described in this invention. Figure 3 This is a block diagram of the bidirectional explosion-proof valve collaborative triggering timing matching control logic of the present invention. Detailed Implementation

[0023] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0024] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0025] Example 1: like Figure 1 As shown, this invention provides an intelligent prediction and triggering control method for pre-explosion signs of an active bidirectional explosion-proof valve, comprising the following steps: S1. Multi-dimensional sensing units are deployed on the forward and reverse pipeline sides of the explosion-proof valve. A sub-microsecond time synchronization network of all sensing nodes is constructed with the edge controller as the main clock node. The inherent response delay calibration of the sensors, the calibration of the baseline operating condition characteristics, and the combined weight calculation of the characteristic parameters are completed, as follows: Sensing array deployment: Equally spaced along the axial direction on the forward pipeline side (upstream of material conveying) of the target bidirectional explosion-proof valve. M Each sensing unit is arranged at equal intervals along the axial direction on the reverse side of the pipeline (downstream of material conveying). N One sensing unit M , N All have ≥3 units, and the axial spacing between adjacent sensing units is 1-3m. Each sensing unit integrates a high-frequency dynamic pressure sensor with a sampling frequency ≥10kHz, an infrared combustible component / dust concentration sensor, an electrostatic sensor, and an armored temperature sensor. One flame precursor infrared sensor with a response time ≤1ms is installed at each of the forward and reverse valve ports of the explosion-proof valve body, forming a multi-physics field sensing network covering the bidirectional pipeline.

[0026] Sub-microsecond time synchronization system construction: An edge controller serves as the master clock node, with all sensing units acting as slave clock nodes. An industrial Ethernet network is established using the IEEE 1588 PTP V2 precision time protocol, achieving sub-microsecond time synchronization within ±100ns for all sensing nodes. A standard step signal generator is used to calibrate the inherent response delay of each sensor type. The calibration method involves synchronously inputting a standard step physical signal to all sensors, measuring the time difference between the rising edge of each sensor's output signal and the standard signal, which represents the inherent response delay of that sensor type. The calibration results are stored in the edge controller's calibration database, providing a benchmark for subsequent phase calibration.

[0027] Baseline calibration: Under stable and normal production conditions, continuously collect full parameter baseline data for a duration T≥30min, and calculate the baseline mean, baseline standard deviation, baseline mean and standard deviation of the rate of change of time, and baseline mean and standard deviation of the spatial gradient for each precursor characteristic parameter.

[0028] Perception parameter combination weight calculation: Combining the subjective weights of the analytic hierarchy process (AHP) and the objective weights of the entropy weight method, the combined weight of each feature parameter is calculated to solve the problem that a single weight cannot take into account both expert experience and field data characteristics. The calculation formula is as follows: In the formula, For the first i The final combined weights of the feature parameters satisfy the following conditions: ; For the first i The subjective weights of each feature parameter are determined by the analytic hierarchy process based on the feature contribution of the explosion precursor. For the first i The objective weights of each feature parameter are calculated by the entropy weight method based on the degree of dispersion of the benchmark working condition data. The total number of precursor characteristic parameters, k≥5, includes 5 core parameters: pressure, concentration, static electricity, temperature, and flame precursor.

[0029] S2. Real-time acquisition of sensing data, completion of phase calibration and adaptive chaotic amplification feature enhancement of weak early-stage explosion precursor signals, calculation of multi-dimensional anomaly degree of single parameter, and then calculation of the explosion precursor coupling hazard degree of forward and reverse pipelines, as detailed below: Real-time data acquisition and phase calibration: Feature parameter data of all sensing units are acquired in real time at a sampling frequency of ≥10kHz. First, phase calibration is performed on the acquired raw signals to correct the phase error caused by the inherent response delay of the sensors and time synchronization deviation. The calibration formula is as follows: In the formula, for tTime of the first m The first sensing unit i The effective signal after phase calibration of each parameter; These are the original values ​​collected by the sensor; For the first i The inherent response delay calibration value of this type of sensor; The real-time time deviation between the slave clock and the master clock of the corresponding sensing unit is positive when the slave clock lags behind the master clock and negative when it leads.

[0030] Adaptive Chaotic Amplification and Feature Enhancement of Weak Precursor Signals: Addressing the issue of weak nanovolt-level precursor signals in the early stages of an explosion being overwhelmed by strong operating noise in the calibrated signal, a Douffing-Holmes chaotic system with adaptive periodic driving force is employed to achieve selective amplification and noise suppression of the weak signal. The processing flow is as follows: Figure 2 As shown, the system dynamic equations are as follows: In the formula, The system damping coefficient is... , The restoring force coefficient is nonlinear. As a periodic reference driving force, For the driving force amplitude, The driving force angular frequency is matched with the characteristic frequency of the precursor signal to be measured. The signal input gain, with a value ranging from 10 to 100; The background is Gaussian white noise.

[0031] The signal enhancement processing logic is as follows: adaptively adjust the driving force amplitude. F This allows the system to operate at the critical threshold between a chaotic state and a large-scale periodic state. When the precursor signal to be measured is input, the system transitions from a chaotic state to a large-scale periodic state, achieving selective amplification of the weak precursor signal matched to the reference frequency and strong noise suppression. The enhanced characteristic signal is output through phase space trajectory extraction. Subsequent anomaly calculations are all performed based on the enhanced feature signals; finally, residual high-frequency noise is removed by 20-50ms moving average filtering, while retaining the low-frequency pulsation and gradient features of the pre-explosion precursor.

[0032] Single-parameter multi-dimensional anomaly calculation: For each feature parameter, three dimensions—absolute value deviation, time rate of change deviation, and spatial propagation gradient deviation—are simultaneously fused to calculate the real-time anomaly degree. This addresses the pain point that a single dimension cannot distinguish between normal operating condition fluctuations and pre-explosion precursors. The calculation formula is as follows: In the formula, for t Time of the first mThe first sensing unit i Real-time anomaly degree of each feature parameter; The rate of change over time, The axial spatial gradient; , The first i The baseline mean and standard deviation of each parameter; , These are the benchmark mean and standard deviation of the rate of change over time, respectively. , These are the baseline mean and standard deviation of the spatial gradient, respectively. , , The weights are the absolute value of the parameter, the temporal gradient, and the spatial gradient, respectively, satisfying... + + =1, determined by standard explosion test; where the pressure parameter spatial gradient weights are... ≥0.5, highlighting the spatial propagation characteristics of the explosion wave.

[0033] Calculation of pre-explosion coupling hazard in bidirectional pipelines: Combining combined weights, anomalies, location weights, and propagation direction coefficients, the pre-explosion coupling hazard of forward and reverse pipelines is calculated separately to quantify the explosion risk level. The formula for calculating the pre-explosion coupling hazard of the forward pipeline is as follows: The formula for calculating the pre-explosion coupling hazard of the corresponding reverse pipeline is as follows: In the formula, , The explosion precursor coupling hazard levels are respectively for forward and reverse pipelines; , These represent the number of forward and reverse sensing units, respectively. and The calculation method and same, For the first positive pipeline n The first sensing unit, the first i Real-time anomaly degree of each parameter For the reverse pipeline n The first sensing unit, the first i Real-time anomaly degree of each parameter; , The first m , n The position weight of each sensing unit is determined by its proximity to the explosion-proof valve; the weight increases as the unit approaches the valve. The formula for calculating the position weight on the forward pipeline is as follows: ,in For the first mThe distance from each sensing unit to the explosion-proof valve This represents the total length of the forward sensing array. The compensation coefficient is the axial spacing between adjacent sensing units; the formula for calculating the position weight of the reverse pipeline is the same as that for the forward pipeline. , These are the propagation characteristic coefficients of the forward and reverse pipeline signals, respectively. The propagation direction is determined by the order in which the abnormal signal peaks of adjacent sensing units appear: in the forward pipeline, if the abnormal peak of the upstream sensing unit appears earlier than that of the downstream (closest to the valve) sensing unit, it is determined that the propagation is towards the explosion-proof valve. =1, otherwise =0.5; In a reverse pipeline, when the abnormal peak value of the downstream sensing unit is earlier than that of the upstream (closer to the valve) sensing unit, it is determined to be propagating towards the explosion-proof valve. =1, otherwise =0.5; This coefficient is used to filter out non-propagating operating condition fluctuations, further reducing the false judgment rate.

[0034] S3. Calculate the real-time operating condition fluctuation coefficient, generate a three-level dynamic adaptive prediction threshold, and complete the graded prediction of explosion risk, as detailed below: Real-time operating condition fluctuation coefficient calculation: Based on the parameter dispersion within the sliding time window, the real-time operating condition fluctuation coefficient is calculated to provide a basis for adjusting the dynamic threshold. The calculation formula is as follows: In the formula, For real-time operating condition fluctuation coefficient, satisfying ≥1, the greater the fluctuation in operating conditions, The larger the value; Sliding time window Inner i Real-time standard deviation of each parameter The value ranges from 100ms to 500ms.

[0035] Dynamic adaptive prediction threshold calculation: Based on the baseline threshold and the real-time operating condition fluctuation coefficient, a three-level dynamic prediction threshold is calculated to solve the problem of poor adaptability of fixed thresholds. The calculation formula is as follows: In the formula, , , These are dynamic thresholds for early warning level, pre-trigger level, and emergency trigger level, respectively. , , The baseline threshold is determined by a standard explosion test. The baseline threshold is taken as 3 times, 5 times, and 10 times the standard deviation of the coupled hazard level under normal operating conditions, satisfying 0 < < < Typical value =2、 =5、 =10.

[0036] Intelligent prediction and classification of pre-explosion signs: Based on the comparison results of coupled hazard level and dynamic threshold, a three-level classification control is implemented: Warning level: When ≥ or ≥ At that time, the on-site audible and visual warning is activated, the sensor sampling frequency is increased to 50kHz, and high-speed data storage is enabled; Pre-trigger level: when ≥ or ≥ When the explosion-proof valve's energy storage mechanism is activated for pre-charging, the valve disc performs a pre-start action, reducing the valve disc closing response time by more than 60%. Emergency Trigger Level: When ≥ or ≥ When the time comes, the coordinated triggering timing control of the two-way explosion-proof valve will be executed immediately.

[0037] S4. Calculate the real-time propagation speed of the explosion wave and the optimal triggering time of the two-way valve, and execute differentiated collaborative triggering control and adaptive valve closing speed control to achieve collaborative protection against two-way explosion-proofing, avoiding the superposition of detonation waves and water hammer effect. The control logic is as follows: Figure 3 As shown, specifically: Real-time propagation velocity calculation of the explosion wave: Based on the time difference of abnormal signals between adjacent sensing units, the real-time propagation velocity of the explosion wave along the pipeline is calculated using the following formula: In the formula, , These represent the propagation speeds of the forward and reverse explosion waves, respectively. The axial spacing between adjacent sensing units; , These represent the propagation time differences between adjacent sensing units for positive and negative abnormal signals, respectively, calculated using a cross-correlation algorithm.

[0038] Calculation of the optimal triggering time for a two-way valve: Based on the propagation speed of the explosion wave and the mechanical response characteristics of the valve, the optimal triggering time that perfectly matches the moment when the valve disc is completely closed and the moment when the explosion wave arrives at the valve orifice is calculated, achieving proactive pre-closing. The formula for calculating the optimal triggering time is as follows: The formula for calculating the optimal trigger time in the opposite direction is as follows: In the formula, , These are the optimal trigger times for both forward and reverse directions; The current moment; , These are the distances from the leading edge of the abnormal signal to the valve port; The rated full closing time of the explosion-proof valve is determined by the valve's mechanical characteristics.

[0039] Exception handling logic: When the calculated optimal trigger time... ≤ or ≤ When the valve is activated, the corresponding explosion-proof valve on the side is immediately triggered, and the valve closing speed is increased to the maximum rated speed.

[0040] Bidirectional Cooperative Triggering Timing Control: Based on bidirectional hazard levels and optimal trigger times, differentiated cooperative control is executed to address the pain point of misaligned action timing of bidirectional valves. One-way hazard triggering: only ≥ At that time, The forward explosion-proof valve is constantly triggered to fully close, while the reverse explosion-proof valve remains in a pre-triggered state; only ≥ At that time, The reverse explosion-proof valve is constantly triggered to close completely, while the forward explosion-proof valve remains in a pre-triggered state to prevent the explosion wave from reflecting and forming a reverse detonation; Two-way danger triggering: when ≥ and ≥ At times, comparison and The explosion-proof valve on the side that triggers earlier will be prioritized, while the valve on the other side will be delayed. Trigger; Set maximum delay threshold =20ms, when calculated > At that time, take = To avoid excessive delays that could lead to protection failure, and to prevent simultaneous closure of both sides from causing a sudden increase in pressure inside the pipeline and a secondary explosion.

[0041] Adaptive control of valve closing speed: Based on the differential control of coupling hazard, the valve closing speed is controlled to avoid the water hammer effect caused by rapid closing. The calculation formula is as follows: In the formula, This refers to the real-time closing speed of the valve disc. This refers to the maximum rated closing speed of the valve disc. The coupling hazard level of the corresponding side pipe.

[0042] S5. Emergency closed-loop control is executed based on the feedback of the explosion-proof effect. The predictive model is adaptively self-corrected through iterative algorithms, which can realize the model's self-learning and improve its reliability throughout its entire life cycle. Specifically: Explosion-proof effect closed-loop feedback control: After the explosion-proof valve is triggered to close, the pressure and flame signal after the valve are monitored in real time. If an abnormal increase in pressure or a flame signal is detected after the valve, the valve disc is immediately triggered to close in an emergency manner, and the plant's SIS system is linked to perform emergency shutdown and inert gas fire extinguishing operations, forming a protection closed loop.

[0043] Predictive model adaptive self-correction: Based on the actual explosion results after triggering, iteratively optimizes the combined weights of feature parameters to solve the problem of accuracy decay in long-term model operation, first clarifying the prediction error. Definition: An actual explosion occurred, triggering the system's emergency response: the prediction was accurate. =0; An explosion actually occurred, but the system did not trigger emergency actions: a missed detection. =1; No explosion actually occurred; the system triggered an emergency response: a false alarm. =-1; No explosion actually occurred, and the system did not trigger emergency actions: the prediction was accurate. =0.

[0044] The weight self-correction iterative formula is as follows: In the formula, The combined weights are updated iteratively; The model learning rate; To predict errors, a value of 1 is used for missed predictions, -1 for incorrect predictions, and 0 for accurate predictions. For the first i Real-time average anomaly of each parameter.

[0045] Weight constraints and normalization: Weights after iteration Upper and lower limits must be satisfied, i.e. Values ​​exceeding the upper and lower limits are taken according to the limits; after iteration, all weights are normalized to ensure... .

[0046] Equipment status self-calibration: Based on the actual closing time of the explosion-proof valve and the valve disc sealing status, it updates in real time. The mechanical characteristic parameters are adjusted to modify the trigger timing calculation model and adapt to the characteristic changes caused by equipment aging; the sensor response delay and time synchronization accuracy are calibrated every 3 months to ensure stable sensing accuracy throughout the entire life cycle.

[0047] S6. Upload the predicted and triggered data to the industrial internet platform in real time, linking the plant's DCS, SIS, and emergency management system to achieve full-process linkage in emergency response; at the same time, store all sensing data, predicted results, triggered actions, and equipment status data to form a full lifecycle database for accident tracing, model optimization, and equipment maintenance.

[0048] Therefore, this invention adopts the above-mentioned intelligent prediction and triggering control method for explosion precursors of an active bidirectional explosion-proof valve. Through sub-microsecond multi-sensor synchronous calibration, chaotic enhancement of weak early-stage explosion precursor signals, intelligent prediction coupled with multi-physics field coupling, and bidirectional explosion-proof collaborative triggering technology, it achieves proactive protection 10-50ms before the explosion occurs. This significantly reduces feature calculation errors and on-site misjudgment rates, improves the reliability of bidirectional explosion-proof protection, and solves the core pain points of existing explosion-proof valves, such as delayed response, difficulty in protecting against reverse detonation, and high misjudgment and missed judgment rates.

[0049] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An explosion precursor intelligent prediction trigger control method of an active bidirectional explosion-proof valve, characterized in that, Includes the following steps: S1. Multi-dimensional sensing units are deployed on the forward and reverse pipeline sides of the explosion-proof valve. A sub-microsecond time synchronization network of all sensing nodes is constructed with the edge controller as the main clock node to complete the calibration of the inherent response delay of the sensor, the calibration of the baseline working condition characteristics, and the combined weight calculation of the characteristic parameters. S2. Real-time acquisition of sensing data, completion of phase calibration and adaptive chaotic amplification feature enhancement of weak early warning signals of explosion, calculation of single-parameter multi-dimensional anomaly degree, and then calculation of the explosion precursor coupling hazard degree of forward and reverse pipelines. S3. Calculate the real-time operating condition fluctuation coefficient, generate a three-level dynamic adaptive prediction threshold, and complete the graded prediction of explosion risk. S4. Calculate the real-time propagation speed of the explosion wave and the optimal triggering time of the bidirectional valve, and execute differentiated collaborative triggering control and adaptive control of valve disc closing speed. S5. Emergency closed-loop control is executed based on the feedback of the explosion-proof effect, and the adaptive self-correction of the prediction model is completed through iterative algorithm; S6. Upload the predicted and triggered data to the industrial internet platform in real time, link the factory emergency system, and complete the data storage and traceability of the entire process.

2. The explosion precursor intelligent pre-judgment trigger control method of the active bidirectional explosion-proof valve according to claim 1, characterized in that: In S1, all sensing units and edge controllers are networked via the IEEE 1588 PTP V2 Precision Time Protocol to achieve sub-microsecond time synchronization within ±100ns for all sensing nodes; the inherent response delay of each type of sensor is calibrated using a standard step signal generator; the combined weights of the feature parameters are calculated using the following formula: ； In the formula, is the final combination weight of the first characteristic parameter, satisfying i ; ; is the subjective weight of the first characteristic parameter, calibrated by the analytic hierarchy process; i ; is the objective weight of the first characteristic parameter, calculated by the entropy weight method; i ; is the total number of precursor characteristic parameters.

3. The explosion precursor intelligent pre-judgment trigger control method of the active bidirectional explosion-proof valve according to claim 2, characterized in that: In S2, phase calibration is first performed on the acquired raw signal. The calibration formula is as follows: ； In the formula, for t Time of the first m The first sensing unit i The effective signal after phase calibration of each parameter; These are the original values ​​collected by the sensor; For the first i The inherent response delay calibration value of this type of sensor; The real-time time deviation between the slave clock and the master clock of the corresponding sensing unit is positive when the slave clock lags behind the master clock and negative when it leads.

4. The method for intelligent prediction and triggering control of explosion precursors in an active bidirectional explosion-proof valve according to claim 3, characterized in that: In S2, the calibrated signal is enhanced using a Dove-Holmes chaotic system with adaptive periodic driving force to enhance the weak precursor signal characteristics. The system dynamic equations are as follows: ； wherein is the system damping coefficient, , is the nonlinear restoring force coefficient, is the periodic reference driving force, is the driving force amplitude, is the driving force angular frequency, is the signal input gain, is the background Gaussian white noise term; By adaptively adjusting the amplitude of the driving force F By operating the system in a chaotic critical state, selective amplification and noise suppression of precursory weak signals are achieved, resulting in the output of enhanced characteristic signals. .

5. The method for intelligent prediction and triggering control of explosion precursors in an active bidirectional explosion-proof valve according to claim 4, characterized in that: In S2, the single-parameter multidimensional anomaly degree is calculated using the following formula: ； In the formula, for t Time of the first m The first sensing unit i Real-time anomaly degree of each feature parameter; For the rate of change over time, The axial spatial gradient; , The first i The baseline mean and standard deviation of each parameter; , These are the benchmark mean and standard deviation of the rate of change over time, respectively. , These are the baseline mean and standard deviation of the spatial gradient, respectively. , , The weights are the absolute value of the parameter, the temporal gradient, and the spatial gradient, respectively, satisfying... + + =1.

6. The explosion precursor intelligent pre-judgment trigger control method of the active bidirectional explosion-proof valve according to claim 5, characterized in that: In S2, the formula for calculating the pre-explosion coupling hazard of the forward pipeline is: ； The formula for calculating the pre-explosion coupling hazard of a reverse pipeline is: ； In the formula, , These represent the risk levels of forward and reverse coupling, respectively. , These represent the number of forward and reverse sensing units, respectively. and The calculation method and same, For the first positive pipeline n The first sensing unit, the first i Real-time anomaly degree of each parameter For the reverse pipeline n The first sensing unit, the first i Real-time anomaly degree of each parameter; , The first m , n The positional weight of each sensing unit is greater the closer it is to the explosion-proof valve; , These are the propagation characteristic coefficients of the forward and reverse pipeline signals, respectively. The coefficient is 1 when the abnormal signal propagates to the explosion-proof valve, and 0.5 otherwise.

7. The explosion precursor intelligent pre-judgment trigger control method of the active bidirectional explosion-proof valve according to claim 6, characterized in that: In S3, the real-time operating condition fluctuation coefficient is calculated using the following formula: ； In the formula, For real-time operating condition fluctuation coefficient, satisfying ≥1; For the first time window within the sliding time window i Real-time standard deviation of each parameter; The three-level dynamic adaptive prediction threshold is calculated using the following formula: ； In the formula, , , These are dynamic thresholds for early warning level, pre-trigger level, and emergency trigger level, respectively. , , As the baseline threshold, satisfying 0 < < < ; when ≥ or ≥ When the warning level is triggered, ≥ or ≥ The pre-trigger level is triggered when ≥ or ≥ Emergency trigger level is triggered at this time.

8. The explosion precursor intelligent pre-judgment trigger control method of the active bidirectional explosion-proof valve according to claim 7, characterized in that: In S4, the real-time propagation speed of the explosion wave is calculated using the following formula: ； In the formula, , These represent the propagation speeds of the forward and reverse explosion waves, respectively. The axial spacing between adjacent sensing units; , These represent the propagation time differences between adjacent sensing units for positive and negative abnormal signals, respectively. The optimal triggering times for the forward and reverse explosion-proof valves are calculated using the following formulas: ； ； In the formula, , These are the optimal trigger times for both forward and reverse directions; The current moment; , These are the distances from the leading edge of the abnormal signal to the valve port; The rated full closing time for the explosion-proof valve.

9. The explosion precursor intelligent pre-judgment trigger control method of the active bidirectional explosion-proof valve according to claim 8, characterized in that: In S4, the valve closing speed is adaptively controlled by the following formula: ； In the formula, This refers to the real-time closing speed of the valve disc. This refers to the maximum rated closing speed of the valve disc. The coupling hazard level of the corresponding side pipe.

10. The method for intelligent prediction and triggering control of pre-explosion signs in an active bidirectional explosion-proof valve according to claim 9, characterized in that: In S5, the self-correcting iterative formula for the prediction model's weights is: ； In the formula, The combined weights are updated iteratively; The model learning rate; To predict errors, a value of 1 is used for missed predictions, -1 for incorrect predictions, and 0 for accurate predictions. For the first i The real-time average anomaly of each parameter; the weights after iteration must satisfy upper and lower bound constraints and undergo normalization to ensure... .

Images